andreacavalli/rocksdb
0
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Assertion failure for L0-L1 compactions.

Browse Source 
Summary:
For level-0 compactions, we try to find if can include more L0 files
in the same compaction run. This causes the 'smallest' and 'largest'
key to get extended to a larger range. But the suceeding call to
ParentRangeInCompaction() was still using the earlier
values of 'smallest' and 'largest',

Because of this bug, a file in L1 can be part of two concurrent
compactions: one L0-L1 compaction and the other L1-L2 compaction.

This should not cause any data loss, but will cause an assertion
failure with debug builds.

Test Plan: make check

Differential Revision: https://reviews.facebook.net/D10677
This commit is contained in:
Dhruba Borthakur 2013-05-08 10:19:24 -07:00
parent 988c20b9f7
commit a8d3aa2c26
1 changed files with 5 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
db/version_set.cc
View File

@ -2082,6 +2082,11 @@ Compaction* VersionSet::PickCompaction() {
// which will include the picked file.
c->inputs_[0].clear();
current_->GetOverlappingInputs(0, &smallest, &largest, &c->inputs_[0]);
// If we include more L0 files in the same compaction run it can
// cause the 'smallest' and 'largest' key to get extended to a
// larger range. So, re-invoke GetRange to get the new key range
GetRange(c->inputs_[0], &smallest, &largest);
if (ParentRangeInCompaction(&smallest, &largest,
level, &c->parent_index_)) {
delete c;